The Impact of Citizens’ Assemblies on Democratic Resilience: Evidence from a Field Experiment

EPSA 2024, Cologne

Tim
Wappenhans

Humboldt-Universität zu Berlin

Bernhard
Clemm

GESIS Leibniz Institute for the Social Sciences

Felix
Hartmann

Copenhagen Business School

Heike
Klüver

Humboldt-Universität zu Berlin

July 1, 2024

Motivation

Anti-democratic sentiment

Antidote

Democratic deliberative events

Preview

What we do

  • preregistered field experiment in Germany
  • citizens’ assemblies w/ MPs
  • match treated (n=435)
    to control units (n=2,675)

What we find

  • large, robust effects on democratic resilience
  • not driven by selection

Expectations

Citizens’ assemblies and democratic resilience


The case

Field experiment

Hallo Bundestag

  • MPs deliberating with
    \(\approx\) 25 constituents
  • 6 electoral districts
  • 17 events
  • 8 hours
  • trained mediators
  • vetted info material
  • national political topics
  • deliberating \(\neq\) deciding

Treatment

Constituents and MP deliberating (Erfurt-Weimar)

Analytical Strategy

Matching

Two-sample approach

  • residence register: legally bound to one specific objective
    • two separately drawn random samples
  • match participants (Xu and Yang 2022)
    • age, education, household size, household income, employment status, past turnout, interest in politics
  • ATT estimated using:

\[ Y_i = \alpha + \beta \text{Treatment}_i + \gamma X_i + \epsilon_i \]

Descriptives


Participation

Covariate balance

Results

Main specification

ATTs

Results for individual items Appendix

Main specification

ATTs

Results for individual items Appendix

Main specification

ATTs

Results for individual items Appendix

Main specification

ATTs

Results for individual items Appendix

Main specification

ATTs

Results for individual items Appendix

Internal valdity

Results driven by selection?

We don’t think so

  • selection
    • exclude 40% of control
  • matching
    • exact matches across
  • resolute sampling
    • sending trained staffers
  • sensitivity Appendix
    • confounder needs to be 11x strong as political interst
  • within-changes Appendix
    • match treated respondents pre/post intervention

Takeaway 🥡

Results of citizens’ assemblies

Main

  • increase trust, efficacy, participation
  • reduce conspiracy thinking
  • longevity (ongoing)

Future

  • spillover
  • backlash

Implications

  • scalability
  • political demand



Hit me up

📬 tim.wappenhans@hu-berlin.de

🌐 timwappenhans.com

🐦 @TimWapps

References

Boulianne, Shelley. 2018. “Mini-Publics and Public Opinion: Two Survey-Based Experiments.” Political Studies 66 (1): 119–36.
———. 2019. “Building Faith in Democracy: Deliberative Events, Political Trust and Efficacy.” Political Studies 67 (1): 4–30.
Gastil, John, E Pierre Deess, Phil Weiser, and Jordan Meade. 2008. “Jury Service and Electoral Participation: A Test of the Participation Hypothesis.” The Journal of Politics 70 (2): 351–67.
Krakowski, Krzysztof, Bernhard Clemm von Hohenberg, and Davide Morisi. 2024. “Does School Debating Reduce Vulnerability to Misinformation? A Field Experiment in Poland.”
Xu, Yiqing, and Eddie Yang. 2022. “Hierarchically Regularized Entropy Balancing.” Political Analysis, 1–8.

Appendix

Trust

Back to Main

Internal efficacy

Back to Main

External efficacy

Back to Main

Conspiracy thinking

Back to Main

Sensitivity analysis

Back to Internal validity

Within-subject design

Back to Internal validity